A CLIMATIC MODEL FOR LOCATION OF PLANT-SPECIES IN FLORIDA, USA

A climatic-envelope model for 125 woody plant species native to Florida was built using data from site floristic lists, species range maps and long-term climatic records for 106 meteorological stations. The model predicts that a species will occur at a site as long as none of the species' climatic limits is exceeded by the local climatic data. The model was tested for its ability to predict species lists for two different sets of sites and for species ranges (i.e. site lists for various species). Median prediction success for these tests varied from 85% to 88% (from 96% to 100% for predictions accurate to within 100 km of range-map distance). This supports the hypothesis that climatic factors, especially winter temperatures, exert important direct and/or indirect control over the natural distribution of woody species in Florida, despite pervasive variations in substrate. The model should be useful for studying climatic factors that may limit species ranges in Florida, as well as potential changes in ranges of species after moderate climate change. Until actual range limitation mechanisms are better known, there may be no practical alternative to simple envelope models for simultaneously treating large numbers of plant taxa and sites. Furthermore, even at local scales, where substrate, topography, fire, competition and other factors become more important, there may be no present alternative to purely climate-based models as a first step in the geographic study of plant-environment relationships.